This paper discusses a new control approach for robotic-assisted tele-echography. The control architecture follows a hierarchical approach, where explicit Cartesian force control arises as the primary task while orientation control is designed in the null space. The robot dynamics is driven by a 3-D time-of-flight camera and a force sensor. Based on depth camera and force data, contact stiffness is anticipated, allowing control adaptation before contact. This approach is adequate for tele-echographic tasks since it reduces robot dynamics before contact, enabling a smooth transition from free space to contact and vice versa. In contact, the environment stiffness is estimated online using force data and the manipulator inertial properties. A strong correlation between the stiffness perceived by the controller and the effective mass exists, being this correlation used in the estimation algorithm to improve force control performance. A set of experiments assess the control architecture, highlighting the relation between perceived stiffness and robot effective mass, including also clinical validation. Note to Practitioners —The teleoperation control architecture described in this paper relies on computed-torque techniques that take into account robot dynamics, contact stiffness estimation, task-space formulation, and null-space design. Our application is on the medical field, where a physician telecontrols a robot to obtain ultrasound images from a patient. The robot exhibits compliant behaviors in Cartesian space (for safety reasons) managed by the task-space formulation, and stiffer behaviors in orientation control managed by null-space design. This approach allows accurate orientation control with compliant positioning which is adequate for this type of teleoperation tasks. Due to free space and contact interactions, the proposed control approach anticipates the contact stiffness relying on vision sensing, adapting the control gains accordingly. While in contact, the perceived stiffness has a strong correlation with the effective mass, which is used together with force data for its estimation. This strategy allows slowing down robot dynamics toward contact, guaranteeing a smooth behavior, entailing also high-quality force tracking while in contact. All control details are described in this paper, allowing easy implementation not only for this specific context, but also for other telemanipulation architectures that require remote delicate handling, without jeopardizing force control and orientation control performances.